Are cannabinoids with cancer immunotherapy contributing to early death? A call for caution and further study

In recent years, cannabis has increasingly been integrated into medical treatments including cancer care due to its potential benefits in symptom management. However, concerns are emerging regarding cannabis’ interaction with cancer immunotherapy, specifically its impact on treatment response rates (RRs).

As both a physician involved in the medicinal cannabis industry and an advocate for patient care, I find the discussion and patient informed consent around this interaction critical. Three key studies on this topic offer preliminary data suggesting a potential reduction in immunotherapy efficacy from cannabis use, though the results remain limited and controversial, warranting caution and additional research.

Objective

This article aims to raise the question of appropriateness of medicinal cannabis during cancer-related immunotherapy.

Discussion

Study overview and findings

The first study, published in 2019 by Taha et al.,¹ examined the immunotherapy radiological RR, progression-free survival (PFS) or overall survival (OS) in patients receiving nivolumab, an anti-PD-1 (programmed cell death-protein 1) immunotherapy, for advanced cancer, including lung cancer, renal cell carcinoma and melanoma. This retrospective analysis compared two groups: those on nivolumab alone (89 patients) and those on nivolumab with concurrent cannabis use (51 patients).

Findings showed a significantly lower RR in patients using cannabis (15.9%) than in those on nivolumab alone (37.5%), with a statistically significant odds ratio of 3.13 (see Figure 1). RR was measured using radiological evaluation of tumours and analysis excluded patients with advanced disease with survival less than 2 months. This suggests that patients using cannabis were approximately three times more likely to have a poor response to immunotherapy. Notably, cannabis use did not significantly impact PFS or OS.

View Figure 1–3.

The second study, published in 2020 by Bar-Sela et al.,² which was prospective, aimed to validate and expand on the Taha et al. study findings with a separate cohort of patients. This study compared 34 patients receiving immunotherapy and using cannabis with 68 patients receiving immunotherapy but not using cannabis. All patients were either receiving immunotherapy as either first- or second-line treatment, either anti-PD-1—pembrolizumab or nivolumab, or ipilimumab and nivolumab—or anti-PD-L1 (programmed cell death-ligand 1)—durvalumab or atezolizumab—without reporting on each immunotherapy treatment per se. The cannabis patients were mostly using inhaled cannabis.

The study reported showing that the median time to tumour progression on radiological imaging for cannabis users was only 3.4 months versus 13.1 months for non-users (95% confidence interval [CI] 1.8–6.0 vs 95% CI 6.0–N/A). Furthermore, the study also reported the median OS for cannabis users was 6.4 months vs a much longer 28.5 months for non-users, with significant Log-Rank tests p=0.0025 and p=0.0009, respectively (95% CI 3.2–9.7 vs 95% CI 15.6–N/A) (see Figure 2). To assess whether ectopic phytocannabinoids in cannabis users affect the endocannabinoid (eCB) system, researchers analysed 28 serum eCB and eCB-like lipids in a smaller cohort of 19 non-users and 17 cannabis users. Only one lipid, 2-oleoyl glycerol, showed a significant difference between users and non-users (p<0.04), suggesting minimal or no lasting impact of cannabis on eCB levels before immunotherapy. In contrast, immunotherapy itself significantly altered several eCB and eCB-like levels, regardless of cannabis use. The study did not demonstrate how serum eCB levels are correlated with local cellular or tissue eCB levels or effects. This study additionally faced critique for data accuracy and methodological transparency, as did the 2019 Taha et al. study.

Comparative analysis and interpretation

Concerns raised by Piper et al.³ in their 2024 publication in the medical journal Cancers identified inconsistencies in data reporting and statistical discrepancies, undermining confidence in both studies’ findings. Re-analyses questioned whether baseline demographic differences between groups could account for the observed outcomes. Piper et al. reported identifying statistically significant differences for the Taha study: between cannabis users and non-users in age (62 years vs 67.7 years), smoking status (56.9% vs 40%), liver metastasis (65% vs 19%) and immunotherapy as a second-line therapy (85% vs 55%). Of the Bar-Sela study, Piper et al. also found that “of 22 statistics in the prospective report, four could not be repeated using the same statistics”, and that there were multiple instances of misreporting or analytical errors and significant differences in cohort demographics.

Despite both the Taha and Bar-Sela studies suggesting a possible negative interaction between cannabis and immunotherapy, their methodologies and findings have invited scrutiny. Factors such as smoking and other lifestyle elements, which heavily influence cancer risk and treatment outcomes, should be carefully considered when assessing cannabis’ role in immunotherapy efficacy and cancer progression, and when interpreting the studies. Piper et al. do not dismiss a possible negative interaction between cannabis and immunotherapy but point towards these as possible confounding factors in both of the small studies.

A third study supporting caution

A recent third study by Hadid et al.,⁴ published in 2024, provides further support for the hypothesis that cannabis may antagonise the effects of immunotherapy. In this Detroit (United States of America) single-institution retrospective cohort study of 105 patients with solid malignancies receiving immune checkpoint inhibitors, cannabis use was primarily in the form of prescribed dronabinol (oral synthetic delta-9-tetrahydrocannabinol). In the Hadid et al. study, cannabis use was associated with significantly worse outcomes in patients receiving immunotherapy, including lower median OS (6.7 vs 17.3 months), reduced PFS (4.8 vs 9.7 months) and a markedly lower disease control rate (10.7% vs 37.7%) compared with non-users (see Figure 3). Cannabis use was reported as mostly 5–10mg oral dronabinol per day (82%), versus recreational or smoked cannabis (14%) and cannabidiol (CBD) use (4%). Patients received a range of immune checkpoint inhibitors, primarily nivolumab (60%) and pembrolizumab (25.7%), with smaller proportions receiving combination nivolumab–ipilimumab (5.7%), atezolizumab (5.7%), durvalumab (1.9%) or avelumab (1.0%). Notably, these negative outcomes were most evident in white patients, raising the possibility of ethnicity-related pharmacogenetic variability in cannabinoid metabolism or immune response. Potentially important confounders—such as performance status, comorbidities and socio-economic status—may have been unaccounted for, and although tobacco use was more common among cannabis users (96% vs 74%) and was included as a covariate, it may still have influenced outcomes. This study reinforces prior findings by Bar-Sela et al. and Taha et al. and further underscores the need for prospective clinical trials to determine whether cannabis use should be contraindicated during immunotherapy.

Interpreting the information

One of the challenges in interpreting these studies is the inherent complexity of cannabis as a therapeutic agent. A variety of studies over the years have shown that cannabis affects the immune system through its active compounds, tetrahydrocannabinol (THC), CBD and the minor cannabinoids, which are known to modulate immune cell functions. Interestingly, in contrast, in some in vitro and in vivo studies cannabinoids are shown to inhibit cancer cell proliferation and metastasis, while also promoting apoptosis and suppressing cancer-related angiogenesis.⁵ However, cannabinoids may interfere with anti-tumour immunotherapy responses through several immunomodulatory mechanisms. Klein reported that THC and CBD can suppress pro-inflammatory cytokines such as IL-2 and IFN-γ while increasing anti-inflammatory cytokines like IL-10, potentially dampening immune activity against tumours.⁶ Cencioni et al. showed that anandamide, an endogenous cannabinoid, has been shown to suppress both the proliferation and the release of pro-inflammatory cytokines from human T cells, primarily through the CB₂ receptor.⁷ Simard et al. highlighted that CB₂ receptors are notably expressed in various immune cells, including eosinophils, B lymphocytes, monocytes and T lymphocytes, and the activation of these receptors has been shown to modulate immune responses, suggesting potential interaction with immunotherapy.⁸ Additionally, Hegde et al. found that THC promotes the expansion of myeloid-derived suppressor cells, which inhibit T cell responses and may support tumour immune evasion.⁹ These examples of immune modulation may theoretically counteract immunotherapy, but definitive conclusions are premature given these studies’ limitations, such as the sample sizes, infancy of the research and potential demographic imbalances.

Conclusion

Caution in clinical practice

These findings hold particular relevance for patients considering, or currently using, medicinal cannabis during immunotherapy. As medical practitioners, we must weigh the benefits of cannabis in managing cancer-related symptoms (such as pain, nausea and anorexia) against its potential to compromise treatment efficacy. While cannabis has a promising role in supportive cancer care, these early findings support careful consideration when patients are concurrently undergoing immunotherapy.

As someone engaged in the medicinal cannabis industry, it is crucial to approach the results of these studies with an open mind. The therapeutic potential of cannabis in oncology is still developing, and its integration with immunotherapy may carry unintended risks. Therefore, clinicians are advised to exercise caution—ensuring patients are well informed of potential interactions is essential until further research provides clear guidance.

My experience with Australian oncology patients using medicinal cannabis is that the majority will use oral medication rather than inhaled cannabis, where inhaled cannabis was predominant in the Taha and Bar-Sela studies. A significant consideration for patients is the link between smoking tobacco and metastatic disease progression and cancer survival.

Future directions and need for rigorous research

The discrepancies in the studies discussed highlight a need for rigorous, well-designed prospective trials. Future studies should aim to control for variables such as cancer type, cannabis dosage, socio-economic status and patient lifestyle factors, particularly tobacco use, which may confound results.

Data transparency is vital to facilitate independent verification and bolster public and scientific trust in the findings. Additionally, larger studies should be commissioned to improve statistical power, and, considering the New Zealand and Australian governments currently fund immunotherapy, regulatory bodies such as Medsafe and the Therapeutic Goods Administration should require pharmaceutical companies to make their findings publicly available.